Improvements on HCM Model for g-function Generation and Illustration of Fluid Temperature Prediction

Ground coupled heat pump (GCHP) is a promising solution to the problem of efficient heating/cooling of residential and commercial buildings. Long term modelling of borehole heat exchanger (BHE) is necessary for better design and operation of GCHP. The definition of the boundary condition at the borehole wall has been of particular interest in the modelling of BHE. Equal and uniform heat flux at all the borehole walls is the simplest boundary condition that is used in most analytical models and in some numerical models as well. Uniform temperature at the borehole wall with an imposed total heat flux is more accurate boundary condition since the boreholes are generally hydraulically connected in parallel and the variation of average fluid temperature is negligible along the depth of the borehole, but it is difficult to implement.This thesis presents improvements to a numerical model that uses a fictitious Highly Conductive Material (HCM) that is embedded in the boreholes is used to implement the uniform temperature boundary condition at the borehole wall. A high variation of heat flux at the ends of the borehole was observed in the HCM model. The presence of borehole resistance between the circulating fluid and the borehole wall means that the uniform fluid temperature boundary condition does not imply uniform temperature at the borehole wall. A thin thermally resistive layer was introduced between the HCM and the borehole wall to represent borehole resistance and a semi-circular part was added at the bottom of the borehole to have even distribution of heat flux. The improved model (enhanced HCM, EHCM ) was validated by comparing the g-function generated by the EHCM model with an analytical solution.The EHCM model was illustrated by simulating the operation of a monitored installation located in Baiona, Spain. The model predicted the daily average fluid temperature with reasonable accuracy in comparison with measured data. The reason for some of the inaccuracy was explained. The EHCM model was compared with other models including a commercial software